AudioGenX: Explainability on Text-to-Audio Generative Models
This addresses the problem of explainability for users of text-to-audio generative models, though it appears incremental as it applies existing XAI concepts to a new domain.
The paper tackles the lack of transparency in how text inputs affect generated audio in text-to-audio models by introducing AudioGenX, an explainable AI method that highlights token importance using factual and counterfactual objectives, with experiments showing its effectiveness in producing faithful explanations.
Text-to-audio generation models (TAG) have achieved significant advances in generating audio conditioned on text descriptions. However, a critical challenge lies in the lack of transparency regarding how each textual input impacts the generated audio. To address this issue, we introduce AudioGenX, an Explainable AI (XAI) method that provides explanations for text-to-audio generation models by highlighting the importance of input tokens. AudioGenX optimizes an Explainer by leveraging factual and counterfactual objective functions to provide faithful explanations at the audio token level. This method offers a detailed and comprehensive understanding of the relationship between text inputs and audio outputs, enhancing both the explainability and trustworthiness of TAG models. Extensive experiments demonstrate the effectiveness of AudioGenX in producing faithful explanations, benchmarked against existing methods using novel evaluation metrics specifically designed for audio generation tasks.